4D validation of dynamic treatment delivery for lung tumors

Janssens, Guillaume [IBA] Geets, Xavier [UCL] Orban de Xivry, Jonathan [UCL] Goossens, Samuel [UCL] Delor, Antoine [UCL] Wanet, Marie [UCL] Lee, John Aldo [UCL] Vynckier, Stefaan [UCL] Grégoire, Vincent [UCL] Sterpin, Edmond [UCL]

ABSTRACT – In the context of dynamic IMRT treatment of lung tumors, it is important to validate treatment planning by simulating dose delivery in presence of motion. In this study, Tomotherapy treatment of lung cancer patients were simulated in 4D using Monte-Carlo simulations and image registration. For all patients, agreement between delivered and planned dose were within tolerances, showing that Tomotherapy is suitable for the treatment of lung tumors. KEY WORDS – Lung cancer, 4D, Tomotherapy, Monte-Carlo Introduction Nowadays, most treatment planning systems perform dose optimization using a single 3DCT image, even though tumors may move during delivery, and more specifically in the lung. To mitigate this issue, 4DCT-based delineation and appropriate margin definition are used to ensure good tumor coverage. However, treatment beam and breathing-induced motions interplay may lead to clinically unacceptable delivered doses, whatever the margin definition technique used [1]. Fortunately, 4DCT allows us to simulate the dose delivery during regular breathing, enabling the validation of the 3DCT planned dose. This is particularly important for hypo-fractionated treatments, as the random errors will not average out during treatment over multiple fractions. In this study, the 4D validation of Tomotherapy treatment planning was performed using Monte-Carlo simulations and non-rigid image registration. Simultaneous integrated boost (SIB) and hypo-fractionated stereotactic (SBRT) treatments were considered. Materials and methods For all patients, target volumes were manually delineated on a contrast enhanced 3DCT and propagated through all phases of a 4DCT using non-rigid image registration, assuming regular breathing ensured through patient coaching (audio and video). This led to the definition of an internal target volume (ITVCT). For SIB treatments, the tumor was automatically segmented on all phases of a 4DPET, leading to the definition of the ITVPET, which defined the boost region. A margin accounting for setup errors was then added to define the planning target volumes (PTVCT and PTVPET), and the treatment planning was performed on the 3DCT with the Tomotherapy TPS. Afterwards, dose was computed with the Monte-Carlo model TomoPen [2]. For every projection, i.e. a discrete gantry angle, simulations were performed on a CT phase selected according to breathing period and initial phase. The dose maps were accumulated using non-rigid deformation and finally compared to the approved treatment plan in order to evaluate the differences in dose distribution induced by motion [3]. The overall flow chart is depicted in Figure 1. Moreover, the interplay effects were distinguished from motion itself by comparing the previous results to the case of an infinitively fast breathing (i.e. all projections on every phase). Results For both SIB and SBRT patients, motion itself proved to be the most important cause of discrepancies between planned and delivered dose distributions. However, generalized equivalent uniform doses (gEUDs [4]) computed on target volumes for all patients showed only small differences (gEUD reduction of maximum 1.5%). Besides, for all patients the dose remained in agreement with medical recommendations, confirming that an ITV-based target volume definition strategy is appropriate to account for motion and that interplay effects are not significant (gEUD reduction of maximum 0.6%). So far, no clinically significant interplay effect occurred for none of the patients simulated, showing that Tomotherapy is suitable for the treatment of lung tumors even in specific treatments such as SIB or SBRT. Conclusion The 4D validation of treatment planning for lung tumors is important, especially for hypo-fractionated treatments and dynamic IMRT treatments. However, for all patients included in this study, coached to ensure regular breathing, 4D Monte Carlo treatment planning showed no significant interplay effect and acceptable tumor coverage. Therefore, as long margins are appropriately defined, TomoTherapy can be used for SIB and SBRT treatments, in spite of the fact that motion management techniques such as gating or tracking are not available.